hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Arun C Murthy (JIRA)" <j...@apache.org>
Subject [jira] Updated: (HADOOP-54) SequenceFile should compress blocks, not individual entries
Date Mon, 21 Aug 2006 17:09:16 GMT
     [ http://issues.apache.org/jira/browse/HADOOP-54?page=all ]

Arun C Murthy updated HADOOP-54:

    Attachment: SequenceFile.20060821.perfomance.txt

Doug - here's another patch incorporating the fix (lazy initialization of the inflater) plus
your other suggestions.

I've also attached results of a quick test run of SequenceFile-v3 v/s SequenceFile-v4 (2 runs
of 10k and 100k records respectively). Within that framework it seems reasonable to assume
that creating one ValueBytes object per-record doesn't show  up blatantly (of course we can
further iterate on this); IAC there is a safety net for extreme corner cases (http://issues.apache.org/jira/browse/HADOOP-54#action_12425953).

PS: The existing TestSequenceFile.java in trunk doesn't pass along the 'compress' flag to
mergeTest - I ran the tests after fixing it; the patch reflects that too. Just fyi.

> SequenceFile should compress blocks, not individual entries
> -----------------------------------------------------------
>                 Key: HADOOP-54
>                 URL: http://issues.apache.org/jira/browse/HADOOP-54
>             Project: Hadoop
>          Issue Type: Improvement
>          Components: io
>    Affects Versions: 0.2.0
>            Reporter: Doug Cutting
>         Assigned To: Arun C Murthy
>             Fix For: 0.6.0
>         Attachments: SequenceFile.20060821.patch, SequenceFile.20060821.perfomance.txt,
SequenceFile.updated.final.patch, SequenceFiles.final.patch, SequenceFiles.patch, SequenceFilesII.patch,
> SequenceFile will optionally compress individual values.  But both compression and performance
would be much better if sequences of keys and values are compressed together.  Sync marks
should only be placed between blocks.  This will require some changes to MapFile too, so that
all file positions stored there are the positions of blocks, not entries within blocks.  Probably
this can be accomplished by adding a getBlockStartPosition() method to SequenceFile.Writer.

This message is automatically generated by JIRA.
If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira


View raw message